Many studies assessing attentional cost while maintaining static or dynamic postures were used as a base of reference for postural control 
. Generally, the postural context is a simple motor task in which the attentional cost required to maintain equilibrium is rather low as compared to many day life sensori-motor situations such as walking, grasping an object on the ground, or moving from a sitting to standing position. When seated in the present experiment, no effect of group on CoP sway or RT were observed ( and and ). This postural condition may be considered as a reference situation in which, whatever the BMI, postural and attentional data are similar.
Mean RT and standard deviation for the two groups in the two postural conditions.
Summary of the CoP data and RT for the two groups in the two postural conditions (mean ± standard deviation) in simple task (ST) ans dual task (DT) conditions.
In the unipedal position, the biomechanical configuration (the “centre of mass height/base of support” ratio is clearly higher than the one observed in a seated posture) placed participants in a vulnerable situation in which personal integrity was clearly at risk. As a result, CoP sway (i.e., range and speed) was obviously greater for the controls and obese patients in contrast to the sitting condition. The increase of CoP range reflected the existence of sudden and severe displacements of the CoP towards the limits of the base of support (). Similarly, the increase of CoP speed reflected the quantity of activity necessary for controlling posture and could be interpreted as a deficit of postural control. These increases were larger for the obese patients than for the control group suggesting that for the control group, postural CoP oscillations were relatively limited and maximal CoP excursions remained at distance of the base of support borderline in comparison to the obese patients. For the obese patients, however, the biomechanical constraints imposed to counteract a same postural disruption and to return to an average CoP position required to generate a muscular torque which magnitude must be higher and/or generated faster than in non-obese subjects 
More specifically, similar CoP oscillations in the simple and dual-tasks conditions were observed for both groups (,
). However, whereas control participants were able to maintain a similar RT performance whatever the postural complexity in the dual-tasks condition, obese patients exhibited much longer RTs (RT increased by over 33% in average) in the unipedal stance as compared to the baseline sitting condition (,
). In other words, for the healthy participants, maintaining a unipedal posture did not seem to require a greater amount of attentional resources suggesting that the neural processes for preserving this erect posture can be regarded as rather automatic. Conversely, for the obese patients, the increased RT observed in the same postural condition reflected a higher attentional cost for controlling the unipedal stance because controlling this complex posture probably required a supra-spinal process. In other words, when it was necessary to control online their posture (as it was the case in the unipedal condition), obese patients exhibited some difficulties. Similar interpretations have been proposed in studies investigating postural control in elderly people 
The exclusive focus of attention on postural stability generally results in larger CoP displacements in a situation in which subjects do not need to focus voluntarily on postural stability 
. Only in situations in which the additional cognitive task is truly complex, postural sway increases 
. These observations were generally made while maintaining a bipedal stance. However, this postural task does not seem complex enough to induce a risk of falling. In the present experiment, no instructions were given to participants regarding the priority dedicated to one or the other task. In obese patients, the absence of postural difference between the single- and dual-tasks conditions and the concomitant RT increase during unipedal stance highlighted the priority these obese patients dedicated to postural control. In the unipedal single-task condition, obese patients were already experiencing critical values of CoP displacement. If the obese patients reduced the attention dedicated to the control of posture during the dual-tasks condition, they would probably fall. Thus, the obese patients preferred to decrease their performance in what they considered to be the so-called “secondary” task (increased RT to auditory stimuli), which did not affect their physical integrity. However, in everyday life there are many situations in which we need to manage different attentional tasks together with balance and/or locomotion. The present results clearly suggested that obese patients are not able to easily perform multitasking as healthy adults do.
From a neurophysiological point of view, we hypothesized that skin stretching resulting from obesity may increase the distance between the cutaneous mechanoreceptors and may thus decrease the discrimination threshold of somato-sensory perception. Additionally, a recent study 
showed that proprioception at knee joint is already altered with this pathology in young obese patients aged 7 to 12. Body schema is built on the basis of multisensory inputs including cutaneous and proprioceptive receptors. With obesity, these receptors may provide altered information to the somato-sensory cortical area altering in turn the obese patients' body schema representation. In addition, it has been shown that the limited physical activity, as generally reported in obese patients, also contributed to the alteration of the body schema 
. To develop muscular responses adapted to the postural constraints, the internal model for action must be based on an appropriate body schema 
. Therefore, we believe that obesity altered the subjects' body schema and internal models necessary for postural control, especially in complex postural tasks.
Plasticity of the neuro-muscular system and appropriate internal models may allow better adapting the neurophysiological constraints to this pathology to answer clinical problems such as balance disorders. Some authors highlighted the benefits of weight loss on postural control improvement 
. Associated to weight loss, a rehabilitation program could be optimized by combining physical practice (Strength training, daily living activities, balance training 
), motor cognitive training (mentally simulated motor action 
or virtual reality 
) and learning to better use the sensory information available (e.g., visual anchoring 
). Combining these ways of rehabilitation could act synergistically to improve postural control in obese patients.